Comparing float pCO₂ profiles in the Southern Ocean to ship data reveals discrepancies

Abstract. The Southern Ocean plays a crucial role in the global carbon cycle. Recently, the utilization of biogeochemical (BGC) Argo float data has provided valuable insights into the uptake and release of carbon dioxide (CO₂) by this region. However, significant uncertainty remains regarding the accuracy of pCO₂ (partial pressure of CO₂) values derived from float data. In this study, we compared pCO₂ estimates obtained from float pH data with those from ship-collected data across the Southern Ocean, employing pCO₂-depth, pCO₂-O₂ and CO₂-O₂ vs saturation plots to assess the degree of agreement between these two datasets. Our findings reveal significant systematic differences. A preliminary analysis, ignoring other factors, found that the float data is consistently higher, on average, than the ship data at equivalent depths and oxygen levels. We tested the hypothesis that inaccurate float pH data or float pCO₂ correction process is the main cause of the pCO₂ difference, by quantifying other factors that could produce systematic differences, including: (i) spatial sampling bias, (ii) seasonal bias, (iii) errors in estimated alkalinity, (iv) errors in carbonate system constants, and (v) higher levels of anthropogenic CO₂ in float data. However, none of the other factors were found to be able to fully account for the discrepancies, suggesting issues with float pH data quality and/or the float pCO₂ correction process. Additional analysis included refinements to ship-based and float-based pCO₂ before intercomparison. Overall, we estimate that, in the Southern Ocean, surface pCO₂ from floats is biased high by, on average, at least 10 μatm.